Dharmottara’s Theory of Apoha1 Kei KATAOKA 1 Frauwallner’s view of Dharmottara’s theory of apoha Essentially Dharmottara has changed nothing of Dharmak¯ırti’s theory of apoha.2 This is Frauwallner’s final evaluation of Dharmottara’s theory of apoha. Dhar- mottara is completely dependent on Dharmak¯ırti.3 And when Dharmottara does deviate from Dharmak¯ırti’s position, it does not concern new ideas but remains a mere reshaping of his predecessor’s view.4 Frauwallner’s view, however, seems to be incompatible with Jayanta’s under- standing.5 Frauwallner noticed it and comments on it with a quotation of the rel- evant passage from the Nyayama¯ njar˜ ¯ı.6 There, Jayanta contrasts two theories of 7 apoha by using Man.d.ana’s terminologies of error (vibhrama). One theory of 1A draft of this paper was read at the 17th Conference of the Association for the Study of the History of Indian Thought, Kyoto University, Dec 25th, 2010. I thank Maria Montenegro and Alex Watson for comments. This work was supported by JSPS KAKENHI Grant Number 23242004. 2Frauwallner 1937:287: “An sehr vielem hat er uberhaupt¨ nichts geandert,¨ und wo er andert,¨ han- delt es sich im Grunde doch um keine wirklich neuen Gedanken, sondern nur um eine Umformung der Gedanken Dharmak¯ırtis, wobei er sogar in Widerspruche¨ gerat¨ und das sorgfaltig¨ ausgeglichene Ebenmaß von Dharmak¯ırtis Gedankengebaude¨ stort.”¨ 3Frauwallner 1937:286–287: “Im wesentlichen ist er doch vollkommen von Dharmak¯ırti abhangig.”¨ 4See note 2. 5Frauwallner’s view is also incompatible with the perspective of Jn˜ana¯ sr´ ¯ımitra and Ratnak¯ırti, who criticize Dhamottara by labeling him as saks¯ . adapohav¯ adin¯ (JNA¯ 203.15) and pratis. edhavadin¯ (RNA¯ 54.4). In other words, Dharmottara’s view is different from Dharmak¯ırti’s from the perspec- tive of Jn˜ana¯ sr´ ¯ımitra and Ratnak¯ırti. For the identification of pratis. edhavadin¯ as Dharmottara, see Akamatsu 1984, 1986. 6Frauwallner 1937:280, n.1: “Den Unterschied zwischen dieser Auffassung und der Lehre Dharmak¯ırtis hat Jayantabhat.t.a gut hervorgehoben (Ny. maj., S. 308, 3-8): so ’yam. nantaro¯ na bahyo¯ ’nya eva ka´scid aropita¯ ak¯ aro¯ vyavr¯ . tticchay¯ ayog¯ ad¯ apoha´sabdartha¯ ucyate it¯ıyam asatkhyativ¯ adagarbh¯ a¯ saran. ih. // atha va¯ vikalpapratibim. bakam. jn˜an¯ ak¯ aram¯ atrakam¯ eva/ tad abahyam¯ api vicitravasan¯ abhedop¯ ahitar¯ upabhedam¯ . bahyavad¯ avabhasam¯ anam¯ . lokayatr¯ am¯ . bi- bharti/ vyavr¯ . tticchay¯ ayog¯ ac¯ ca tad apoha iti vyavahriyate seyam atm¯ akhy¯ atigarbh¯ a¯ saran. ih. // Auch Dharmak¯ırti hatte nicht das Vorstellungsbild an sich als Gegenstand der Vorstellungen bezeichnet, sondern nur insofern es nach außen verlegt erscheint. Aber Dharmottara hat die Verschiedenheit von der Erkenntnis als solcher scharfer¨ hervorgehoben, wenn er auch, wie seine Darstellung zeigt, den Zusammenhang zwischen Erkenntnisbild und Gegenstand der Vorstellung nie aus dem Auge ver- loren hat. Uber¨ die tatsachlichen¨ Verhaltnisse¨ bestand also keine Meinungsverschiedenheit. Nur in der Art, wie sie aufgefaßt und ausgedruckt¨ wurden, liegt der Unterschied.” 7 See VV v.1: atmakhy¯ atir¯ asatkhyatir¯ akhyatih¯ . khyatir¯ anyatha/¯ par¯ıks. akan¯ . am¯ . vibhrantau¯ vivad¯ at¯ sa¯ vivicyate// Schmithausen 1965:53: “Da die Philosophen hinsichtlich des Irrtums ver- schiedene Meinungen haben — Selbsterscheinen, Erscheinen von Nichtseiendem, Nichterscheinen und Anderserscheinen —, wollen wir ihm eine Untersuchung widmen.” 116 Journal of Indological Studies, Nos. 24 & 25 (2012–2013) apoha, according to Jayanta, is parallel to the theory of asatkhyati¯ and the other atmakhy¯ ati¯ .8 As Frauwallner correctly assumes, the former can be ascribed to Dharmottara9 and the latter to Dharmak¯ırti,10 although Jayanta never mentions them by name. Frauwallner initially agrees that Jayanta intends to show a the- oretical difference between Dharmak¯ırti’s and Dharmottara’s theories of apoha. Implicitly dismissing Jayanta’s observation, however, Frauwallner concludes that there is no difference of opinion between Dharmak¯ırti and Dharmottara; and that the difference lies only in the manner of understanding and the mode of expres- sion. Again, Frauwallner finds no fundamental difference between Dharmak¯ırti’s and Dharmottara’s theories of apoha.11 But is it the case that Dharmottara’s theory of apoha is essentially non-different from Dharmak¯ırti’s? Can one conclude, as Frauwallner does, that the difference is superficial, not fundamental? Is Jayanta’s observation mistaken or not to be taken seriously? In the following the present author elucidate this problem by carefully examining the models of apoha theory that Dharmak¯ırti and Dharmottara presuppose. 2 Dignaga’s¯ theory of apoha Before examining Dharmak¯ırti’s theory of apoha, we must first review Dignaga’s¯ theory of apoha,12 since Dharmak¯ırti’s model developed out of Dignaga’s.¯ Dignaga¯ replaces universals (jati¯ ) accepted in brahmanical schools with exclu- sion of others (anyapoha¯ ). The word “cow” expresses “not being non-cows”, i.e. the exclusion of non-cows. Universals such as cowness do not exist at all in the external world. It is precisely the exclusion of non-cows that we conceptualize as “cow.” As Jayanta observes, Kumarila¯ understands this exclusion to be equivalent to nonexistence (abhava¯ ), a kind of external reality (vastu) in Kumarila’s¯ system of ontology.13 Kumarila’s¯ interpretation can be justified. According to Dignaga,¯ “cow” de- notes negation of non-cows. Therefore, the meaning of “cow” is precisely the 8See 3.1 and 3.2 of the Nyayama¯ njar˜ ¯ı, Kataoka 2009:34–35. According to the theory of § § atmakhy¯ ati¯ , it is cognition itself (atman¯ ) that appears (khyati¯ ) as an external, real entity, whereas it is a nonexistent (asat), i.e. an unreal, fabricated fiction (aropita¯ ), that appears and that is not dis- tinguished from an external, real entity according to the theory of asatkhyati¯ . Dharmottara’s notion of aropita¯ has a connotation of being unreal, fictional and fabricated. See JNA¯ 230.4: aropitam¯ ity api kalpitam evocyate. “Furthermore, when you say aropita¯ , you refer to precisely kalpita.” 9This is proved by Cakradhara’s commentary NMGBh (132.24: jn˜an¯ arth¯ abhy¯ am¯ anya eveti dharmottarah. ). 10I discussed it in Kataoka 2009:9–12. 11Hattori also, though admitting that Dharmottara “was an original thinker”, and “made some amendments”, maintained that he “basically followed Dharmak¯ırti’s views” (Hattori: 2006:68). 12For Dignaga’s¯ theory of apoha, see Katsura 1988, 1999, Hattori 2006 and Pind 2009. 13Nyayama¯ njar˜ ¯ı, Kataoka 2009:26.3–4: apoho yady abhav¯ atm¯ a¯ bahir abhyupagamyate/ tato bha- vati bhavatkam¯ . vagj¯ alam¯ . na tv asau tatha//¯ “If [we] accepted apoha to be an external negation by nature, the net of [critical] words from you would apply. But it is not so.” Dharmottara’s Theory of Apoha (Kei KATAOKA) 117 nonexistence (abhava¯ ) of non-cows. It can be classified as mutual nonexistence (anyonyabh¯ ava¯ ) among the fourfold classification of nonexistence that Kumarila¯ presents.14 In this way, the nonexistence of non-cows, in place of universal cow- ness, functions as a generic property of all individual cows. As Dignaga¯ describes, “a word denotes things exclusively as qualified by preclusion of other referents 15 (sabdo´ ’rthantaranivr¯ . ttivisis´ ..tan¯ eva bhav¯ an¯ aha¯ ).” Dignaga¯ criticizes the Naiyayika¯ model of tadvat, “[a locus] which has that”: a word expresses individuals qualified by a universal. Dignaga’s¯ own model, how- ever, remains similar to the Naiyayika¯ model of tadvat as a whole. Only the role of universal is replaced by apoha. 3 Dharmak¯ırti’s theory of apoha 3.1 Criticism by Kumarila¯ Assuming Dignaga’s¯ apoha to be equivalent to nonexistence, i.e. a kind of external entity, Kumarila¯ points out various problems with this model.16 As Kumarila¯ points out in the opening verse of his criticism, negation of non-cows, “not being anything that is not a cow,” necessarily presupposes affirmation of cowness. “Clearly those people who postulate that [a word] expresses negation of non-cows as a common property [in fact] have referred precisely to the real entity cowness with the term ‘the exclusion of non-cows’.”17 Indeed double negation leads to many undesirable consequences. Not non-cows amounts to being the same as cow. These and other problems that Kumarila¯ has pointed out in the Slokav´ arttika¯ are explained in prose by Jayanta.18 3.2 Dharmak¯ırti’s strategy Quite correctly Kumarila¯ understands Dignaga’s¯ theory as being essentially paral- lel to the Naiyayika¯ view of tadvat. In order to avoid Kumarila’s¯ criticism, Dhar- mak¯ırti has changed the model drastically.19 He agrees that the Buddhist model of apoha would lead to problems if it were identified with the view of tadvat. But this is not the case, replies Dharmak¯ırti.20 Opponents misunderstand the Buddhist 14SV´ abhava¯ , 2cd–4. 15PSV ad V 36d (quoted in PVSV 62.27-63.1), translation quoted from Pind 2009:106. Cf. Pind 1999. 16For Kumarila’s¯ criticism of apoha, see Hattori 1973, 1975. 17 SV´ apoha, 1: agonivr. ttih. sam¯ anyam¯ . vacyam¯ . yaih. parikalpitam/ gotvam. vastv eva tair uktam ago’pohagira¯ sphut.am// 18See the synopsis of the Nyayama¯ njar˜ ¯ı, Kataoka 2008:17–18. Cf. also Hattori 2006. 19For Dharmak¯ırti’s theory of apoha, see Frauwallner 1932, 1933, 1935; Vetter 1964; Dunne 2004, 2011; and Tillemans 2011. 20 PV I 64 (34.17-18): tenany¯ apohavis¯ . aye tadvatpaks. opavarn. anam/ pratyakhy¯ atam¯ . pr. thaktve hi syad¯ dos. ojatitadvatoh¯ . // “Therefore, concerning the object (i.e. locus) of exclusion of other things, the [opponent’s] explanation of the [Buddhist] view of tadvat (a locus qualified by apoha) is refuted, for if [apoha and its locus were] different, the problem that applies to [the non-Buddhist view of] 118 Journal of Indological Studies, Nos. 24 & 25 (2012–2013) theory, which does not posit negation as a common property from the beginning.